Five locations. Dozens of hours. One story we could finally see clearly.
S21 CCTV
Disclosure notice: This is a true account of a frontline investigation – written in collaboration with Semantics 21. This may contain accounts that some readers find upsetting.
This case study is based on a real experience shared by a law enforcement/investigation agency professional and written in collaboration with Semantics 21. It is presented in a first-person format to reflect the original voice and lived reality of the investigator, with all identifying information removed or adapted in accordance with UK GDPR and safeguarding standards.
Introduction
We were dealing with one of the recent riot events in the UK — fast-moving, chaotic and extremely sensitive. Multiple locations were affected, with serious allegations of damage to public property, transport infrastructure and emergency vehicles.
We had incoming footage from local businesses, housing estates and static city cameras. There were overlapping timelines, conflicting reports and mounting pressure. One suspect claimed to be an innocent bystander, “just out looking for his dog.”
What we needed was clarity and what we didn’t have was time.
S21 CCTV didn’t just help us find the footage. It helped us prove the lie.
We started with the redaction challenge. Footage destined for CPS, partner agencies and disclosure bundles had to be reviewed and prepared immediately. Using facial redaction, S21 CCTV redacted every face while leaving the suspect’s visible — across hours of riot footage, in formats many tools don’t support.
From there, we used the face grouping to follow the suspect across five locations. It revealed:
– One shot with his mask off
– Another with his hood down
– Multiple clips where he appeared in key scenes with no sign of the dog he claimed to be looking for
In one clip, he was seen throwing an object at a bus shelter. In another, climbing onto a police van. Everything was time-stamped and searchable.
From Claim to Timeline
We used auto-redaction to prepare footage for CPS instantly. Facial grouping connected one individual across scenes and confirmed his presence in all five. Each clip was export-ready and admissible.
One excuse collapsed under the weight of the timeline.
Instead of a vague denial, we had sequence, location, action and intent. The final playback sequence was presented in interview. The suspect pled guilty within half an hour.
“We honestly didn’t expect it to work this well. We thought we were getting a faster player. What we got was a whole new way to investigate.”
What if we hadn’t acted?
What if we’d settled for the timeline he gave us? Or blurred every face manually? What if we’d missed the one frame that showed him on top of the van?
It would’ve taken days to stitch it together, weeks to redact and we’d still have doubts.
S21 CCTV didn’t just help us catch him. It helped us see what really happened.
Because of that, we had the proof — clean, clear, and court-ready.
S21 solutions mentioned
S21 CCTV
AI-enhanced CCTV review, enhancement and reporting, designed for frontline teams and forensic labs alike.
Have a story of your own? Whether you’d like to be credited or remain anonymous, we’d love to hear from you. Your experience could help rescue a victim, safeguard a child, or take a dangerous offender off the streets. Please get in touch — your words could make a difference.
Please complete the form with valid company or agency information, including a company or agency-issued email address. We will need to confirm your credentials before issuing a free trial licence.
